Is AI supposed to think for us?
With the rapid rise of generative AI tools like ChatGPT, we can get answers faster than ever before.
Write this for me.
Come up with an idea.
Tell me how to solve this problem.
Which option should I choose?
Type in a question, wait a few seconds, and AI gives you an answer that sounds polished, confident, and often remarkably convincing.
That is incredibly useful.
But I think it also creates one of the biggest traps of the AI era:
We are becoming too eager to ask AI for the “right answer.”
AI should not simply become a machine that thinks so we don’t have to.
Its greatest value may be something far more powerful:
expanding our ability to think.
A Confident AI Answer Is Not the Same as a Correct Answer
There is one thing we need to accept before anything else:
AI can be wrong.
And sometimes, it can be wrong beautifully.
It can produce an answer that is clear, logical, articulate, and completely convincing — while still being based on incorrect information, flawed assumptions, or missing context.
More importantly, the answers AI gives us are heavily influenced by the assumptions we give it.
Imagine there is a problem at work.
You ask:
“I think the problem is poor management. What do you think?”
Now imagine asking:
“I think the problem is the attitude of the frontline staff. What do you think?”
Those two questions may lead the AI down very different paths.
A capable AI may challenge your assumptions, but the basic reality remains: its response is shaped by the information, context, framing, and assumptions you provide.
That is why this mindset is dangerous:
“AI said it, so it must be true.”
AI is not a judge handing down a final verdict.
It is a tool that gives us material to think with.
And that distinction matters.
Start With Your Own Thinking
So how should we use AI?
For me, one principle comes before everything else:
Bring your own thought to the conversation first.
It doesn’t have to be brilliant.
It doesn’t have to be expert-level.
It doesn’t even have to be well organized.
You might simply think:
“Something about this doesn’t feel right.”
That’s enough.
The important thing is to have some kind of starting point that belongs to you.
A hypothesis.
An intuition.
A doubt.
An opinion.
A question.
For example:
“I think the real problem in this organization may be the lack of communication between management and frontline employees.”
Maybe you’re right.
Maybe you’re completely wrong.
At this stage, that doesn’t matter.
Take that thought and give it to AI.
Ask:
“This is how I see the situation. Evaluate my reasoning objectively.”
Now something interesting can happen.
AI may reveal a weakness you hadn’t noticed.
It may introduce a factor you hadn’t considered.
It may challenge the way you framed the problem in the first place.
And that is where AI becomes more than an answer machine.
It becomes a thinking partner.
Use AI to Expand Your “Thinking Space”
I find it useful to imagine thinking as a three-dimensional space.
There are three axes:
Altitude × Perspective × Depth
Or, more simply:
Up and down. Side to side. Deeper in.
AI can help us move along all three.
And when those three dimensions expand, the space in which we are able to think expands with them.
1. Altitude — Change the Level You Are Thinking From
The first axis is altitude.
In other words:
How high above the problem are you looking from?
Take a workplace problem.
You can look at it as a frontline employee.
Then as a team leader.
Then as a manager.
Then as an executive.
Then from the perspective of the entire industry.
And eventually, perhaps, from the perspective of society itself.
The same situation can look completely different depending on where you stand.
A frontline employee might think:
“Why is management forcing us to follow this ridiculous procedure?”
An executive might see the same procedure and think:
“We need this to reduce legal exposure and prevent serious accidents.”
Neither perspective automatically tells us the whole truth.
And the reverse can happen too.
Something that looks perfectly rational in a boardroom may be almost impossible to execute on the ground.
This is where AI becomes useful.
Ask:
“How would a frontline employee see this?”
Then:
“Now analyze it as the CEO.”
Then:
“What does this look like at the industry level?”
You are moving your thinking vertically.
You are changing your altitude.
2. Perspective — Look From More Than One Direction
The second axis is perspective.
This means asking:
From whose position — or through which lens — am I looking at the problem?
Consider the same issue from the perspective of:
a customer,
an employee,
a manager,
an investor,
a competitor,
a family member,
or a regulator.
Then change the intellectual lens.
What does psychology say?
What about economics?
Behavioral science?
Ethics?
Law?
Organizational theory?
A single problem can look completely different depending on the lens through which you examine it.
The limitation is obvious: no individual human being has deep expertise in every field or experience from every possible position.
AI can help us explore beyond those boundaries.
Ask:
“How would someone who strongly disagrees with me see this?”
Or:
“Analyze this from a behavioral economics perspective.”
Or:
“Which stakeholder am I forgetting?”
One of AI’s most powerful abilities is not simply producing information.
It is switching perspectives at extraordinary speed.
3. Depth — Go Beneath the Surface
The third axis is depth.
Problems usually have a visible surface and an invisible structure underneath.
Imagine a workplace where employees constantly criticize one another.
You could ask:
“How can we help everyone get along better?”
You might get some useful advice.
But perhaps you are treating the symptom rather than the cause.
So go deeper.
Why are employees criticizing each other?
Why does that behavior keep repeating?
What incentives reward it?
Are responsibilities unclear?
Is the evaluation system creating competition?
Is information being distributed unevenly?
Is this actually an interpersonal problem — or is it a structural problem disguised as one?
Then go one level deeper still:
“Is the thing I am calling a problem actually the real problem?”
That question alone can completely change the conversation.
AI allows us to keep asking “why?” without getting tired.
What happened?
Why did it happen?
What caused that?
What structure allowed that cause to emerge?
What assumptions support that structure?
And finally:
What if my original assumption was wrong?
That is depth.
Don’t Turn AI Into a Machine That Agrees With You
There is another danger.
When you discuss your ideas with AI, it can sometimes give you very satisfying responses.
“Your reasoning makes sense.”
“That is a valid perspective.”
“Your argument is well founded.”
It feels good.
And that is exactly why we should be careful.
Don’t turn AI into a machine that tells you you’re right.
Use it for the opposite purpose.
Ask AI to attack your thinking.
“What is the weakest part of my argument?”
“Give me the strongest counterargument.”
“Assume my hypothesis is wrong. What else could explain this?”
“What assumptions am I making without realizing it?”
“Take the other person’s side and criticize my position.”
This may be one of the most valuable ways to use AI.
Don’t just use AI to strengthen your ideas.
Let it break them first.
If an idea survives serious criticism, what remains is usually much stronger than the thought you started with.
The Most Important AI Skill Isn’t Writing the Perfect Prompt
People often say that “prompting” is one of the essential skills of the AI era.
I mostly agree.
But I would change the definition.
The most important skill is not being able to write the perfect prompt on your first attempt.
Your first question can be messy.
It can be:
“Something about this seems wrong. Help me figure out why.”
That’s fine.
Read the response.
Then say:
“No, that’s not quite what I mean.”
Then:
“Look at it from management’s perspective.”
Then:
“Now argue against me.”
Then:
“Wait. Is the assumption behind my original question itself wrong?”
The question evolves.
And as the question evolves, your thinking evolves with it.
So perhaps the real AI skill is not simply prompt engineering.
It is question refinement.
Or even more simply:
the ability to ask better questions over time.
Great AI users do not necessarily begin with great questions.
They become better at asking them.
A Six-Step Framework for Thinking With AI
I summarize this approach with six steps:
Think → Ask → Challenge → Expand → Verify → Decide
Think
Start with your own thought.
Even if it is incomplete.
Even if it is wrong.
Give yourself a starting point.
Ask
Bring that thought to AI.
Ask it to analyze, organize, explain, or evaluate your reasoning.
Challenge
Now attack the idea.
Ask for weaknesses.
Ask for counterarguments.
Ask what you may have misunderstood.
Don’t use AI only as an ally.
Sometimes its most valuable role is to become your opponent.
Expand
Move through the three dimensions:
Altitude × Perspective × Depth.
Look from higher up.
Look from another side.
Dig deeper.
Expand the space in which you are thinking.
Verify
Now become skeptical.
Ask:
What is fact?
What is inference?
What is speculation?
What evidence supports this?
Can I verify it through primary sources?
Do other reliable sources agree?
AI can generate persuasive language.
Persuasive is not the same as true.
Decide
Finally, make the decision yourself.
There is nothing wrong with asking AI:
“What would you do?”
But there is a huge difference between considering AI’s recommendation and saying:
“I’m doing this because AI told me to.”
AI does not live your life.
It does not work in your organization.
It does not carry your relationships.
And it does not bear the consequences of your decisions.
Judgment and responsibility still belong to you.
Will AI Make Us Stop Thinking?
There is a growing concern that AI will make people think less.
I think that concern is half right.
If we constantly ask:
“Give me the answer.”
“Write everything for me.”
“Tell me which choice is correct.”
then yes — we may gradually outsource more and more of our thinking.
But reverse the relationship.
Say:
“Here is what I think. Find the holes in it.”
“Show me another perspective.”
“Take me to a higher level of analysis.”
“Go deeper.”
“Challenge the assumption behind my question.”
Now AI is doing something completely different.
It is no longer replacing thought.
It is provoking thought.
Instead of shrinking the territory your mind explores, AI can expand it.
Instead of becoming an intellectual crutch, it can become an intellectual sparring partner.
That difference is not determined by the AI.
It is determined by how we choose to use it.
Don’t Outsource Your Thinking. Expand It.
I don’t think the defining divide of the AI era will simply be between people who “use AI” and people who don’t.
A more important distinction may be this:
Do you know what to outsource to AI — and what must remain yours?
Bring your own idea.
Give it to AI.
Challenge it.
Change altitude.
Change perspective.
Go deeper.
Verify what you learn.
Then make the final decision yourself.
AI does not have to be the end of the thinking process.
It can be the beginning of a much better one.
So I don’t want AI to give me every answer.
I want it to show me the questions I haven’t asked yet.
I want it to challenge the assumptions I didn’t realize I was making.
I want it to show me perspectives I couldn’t see on my own.
And when the conversation ends, I want to be able to look at the same problem from a slightly higher altitude, from more directions, and at a greater depth than when I started.
That, to me, is one of the most valuable ways we can use AI.
AI shouldn’t replace your thinking.
It should expand it.
